I made this to stylize the anime art quality of Flux a bit more. Flux is fairly okay on it's own, and I find I enjoy the flat illustrative art a bit more than the bombastic nature of some of the other lora's out there. So I tried to train a model on some more stylistic character art without all the crazy backgrounds and visual effects. So far, I'm pretty happy with the end result. Model is trained on 130 different images, majority of which were cat girls - hence the name of the Lora. It handles non-catgirl-related prompts just fine for those who might ask. --------------------------------------------------------------------------------------------------- The particulars: Trigger word is kittygal, but you may want to customize the Lora in your Web UI settings and set your own trigger word or just use &lt;lora:KittyGal:1&gt; somewhere in your prompt. I find the model works best at a weight of 1.

Files with SHA256: 8d255a66a80e79b494e12a5ad304ae91703f86b99cb3b25f3d0ce721f8ce66b3